Point-Voxel Based Geometry-Adaptive Network for 3D Point Cloud Analysis  

作　　者：Tian-Meng Zhao Hui Zeng Bao-Qing Zhang Hong-Min Liu Bin Fan 赵天孟;曾慧;张保庆;刘红敏;樊彬(Beijing Engineering Research Center of Industrial Spectrum Imaging,School of Automation and Electrial Engineering University of Science and Technology Beijing,Beijing 100083,China;Shunde Innovation School,University of Science and Technology Beijing,Foshan 528399,China;Beijing Institute of Electronic System Engineering,Beijing 100854,China;School of Intelligence Science and Technology,University of Science and Technology Beijing,Beijing 100083,China;Institute of Artificial Intelligence,University of Science and Technology Beijing,Beijing 100083,China)

机构地区：[1]Beijing Engineering Research Center of Industrial Spectrum Imaging,School of Automation and Electrial Engineering University of Science and Technology Beijing,Beijing 100083,China [2]Shunde Innovation School,University of Science and Technology Beijing,Foshan 528399,China [3]Beijing Institute of Electronic System Engineering,Beijing 100854,China [4]School of Intelligence Science and Technology,University of Science and Technology Beijing,Beijing 100083,China [5]Institute of Artificial Intelligence,University of Science and Technology Beijing,Beijing 100083,China

出　　处：《Journal of Computer Science & Technology》2024年第5期1167-1179,共13页计算机科学技术学报（英文版）

基　　金：supported by the National Natural Science Foundation of China under Grant Nos.62273034,61973029,and 62076026;the Scientific and Technological Innovation Foundation of Foshan under Grant No.BK21BF004.

摘　　要：Point cloud analysis is challenging because of the unordered and irregular data structure of point clouds.To describe geometric information in point clouds,existing methods mainly use convolution,graph,and attention operations to construct sophisticated local aggregation operators.These operators work well in extracting local information but bring unfavorable inference latency due to high computation complexity.To solve the above problem,this paper presents a novel point-voxel based geometry-adaptive network(PVGANet),which combines multiple representations of point and voxel to describe the point cloud from different granularities and can obtain features of different scales effectively.To extract fine-grained geometric features,we design the position-adaptive pooling operator,which uses point pairs’relative position and feature similarity to weight and aggregate the point features at local areas of point clouds.To extract coarse-grained local features,we design a depth-wise convolution operator,which conducts the depth-wise convolution on voxel grids.With an easy addition,fine-grained geometric and coarse-grained local features can be fused,and we can use the geometry-adaptive fused features to complete the efficient shape analysis of point clouds,such as shape classification and part segmentation.Extensive experiments on ModelNet40,ScanObjectNN,and ShapeNet Part benchmarks demonstrate that our PVGANet achieves competitive performance compared with the related methods.

关 键 词：point cloud geometric feature multi-representation part segmentation shape classification 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]